NASA Technical Reports Server (NTRS) 19980237250: Cryogenic Optical Performance of the Cassini Composite InfraRed Spectrometer (CIRS) Flight Telescope

18 Pages·0.71 MB·English

by  NASA Technical Reports Server (NTRS)

Cryogenic optical performance of the Cassini Composite
InfraRed Spectrometer (CIRS) flight telescope
l/v-"/+
f
Patricia Losch, James J. Lyons III, John Hagopian
Optics Branch Code 551
NASA/Goddard Space Flight Center
Greenbelt, MD 20771
Abstract
The CIRS half-meter diameter beryllium flight telescope's optical performance was tested at the instrument
operating temperature of 170 Kelvin. The telescope components were designed at Goddard Space Flight
Center (GSFC) but fabricated out of house and then assembled, aligned and tested upon receipt at GSFC. A
24 inch aperture cryogenic test facility utilizing a 1024 x 1024 CCD array was developed at GSFC
specifically for this test. The telescope's image quality (measured as encircled energy), boresight stability
and focus stability were measured. The gold coated beryllium design exceeded the cold image performance
requirement of 80% encircled energy within a460 micron diameter circle.
Keywords: cryogenic, optical test, encircled energy, beryllium cassegrain telescope
1.0 Introduction
The CIRS is a remote sensor built as an international collaboration at the Goddard Space Flight Center t.
CIRS was launched on NASA's Cassini Orbiter to Saturn in October 1997. Infrared spectroscopy of
emissions from Saturn's and Titan's atmospheres, rings, and surfaces will provide 3-D and temporal
variations of temperature distribution, gas abundances, aerosols, and clouds. Global mapping of Saturn's
rings and icy satellite for composition and thermal properties will also be p_rformed. _
CIRS consists of a 0.5 meter F/6 Cassegrain beryllium telescope, a reference interferometer to provide scan
mechanism control and timing for data sampling, and two science Fourier transform spectrometers: a Mid-
Infrared interferometer and a Far infrared interferometer. The MIR is aclassical Michelson interferometer
with a 7 - 17micron spectral range measured with two linear HgCdTe arrays (FP3 and FP4) consisting of
ten 200 micron square pixels each viewing 0.273 mrad field. The FIR is a polarizing Michelson
interferometer measuring from 17 - 1000 microns using two 4.3 mrad thermopile detectors (FPI). The
CIRS fields of view are shown in Figure 1.
4.300 mrad --_1
O.273mracl
FP4 FP3
Iq'_Pl 1.216 mrad
Figure 1. CIRS fields of view consisting of focal planes 1 (FIR), 3, and 4 (MIR)
CIRS is a passively cooled cryogenic instrument. All subsystems with the exception of the MIR focal plane
operate at 170 Kelvin. The MIR focal plane operates at 80 Kelvin and iscooled using a honeycomb radiative
patch.
' VI_ !J#PI,),M_FN: ',Ir'iAT,P
\
\
..f
L
SUPPORTrUSE
PRIMARY MIRROR /- RAC,IATCR/SU NSHADE
\\
SOLAR BLOCKING
EN[RANCE APERTURE
& ROOF-TOP MIRROR
TELESCOPE MOUNT TUBE FIR POLARIZING
FOLD MIRROR
MIR FOLD MIRROR
FIR COLUMATOR
FIR FOCUS MIRROR
MIR FIXED REFERENCE L_ FIR POLARIZING ROOF MIRROR
RETROREFLECTOR INTERFEROMETER SCAN MECHANISM BEAMSRLITTER FIR FIXED
80K COOLER/FP3&FP4
Figure 2: CIRS Optical Layout
The optical design of CIRS is described by Maymon et aP. A brief description of the optical layout of CIRS
shown in Figure 2 is presented here. Light is collected across the 12.426 mrad field of view by the 0.5 meter, F/6
Cassegrain telescope. The beryllium telescope is mounted via astress-free transition tube to the aluminum optics
bench which houses the collimating optics and interferometers. A parabolic Entrance Aperture Plate (EAP) at the
telescope focal surface serves to block stray light and is mounted in front of and to the field-split mirror. The EAP
contains two apertures coincident with the MIR and FIR fields of view. The MIR and FIR fields of view are
divided by the field split mirror which consists of two weak spherical mirrors each at 45 ° from the telescope
optical axis. The FIR surface of the field split mirror was roughened and works in conjunction with a solar
blocking filter to prevent solar radiation from traveling through the FIR interferometer and damaging the
thermopile detectors. Both the MIR and FIR beams are then collimated by off-axis parabolas and steered into the
interferometers via a fold flat in the MIR and an input polarizer for the FIR. The MIR
contains a Potassium Bromide beamsplitter and compensator. The FIR utilizes a polarizing grid identical to the
input polarizer (but rotated 45°)as its beamsplitter 4. The fixed and moving mirrors are not typical flat mirrors but
are cube corners in the MIR and dihedral mirrors in the FIR. These were developed at GSFC 5to provide relief in
the performance requirements of the mirror scanning mechanism. The dihedrals also serve to rotate the beams'
polarizations. After passing through the fixed and moving legs of interferometer the MIR beam is focused onto
the detector arrays by an F/I germanium lens. The FIR beam is focused by an F/I on-axis parabola through an
output analyzer and onto the corresponding thermopile detector 6. Both signals are then Fourier transformed to
provide spectral information.
2.0 Telescope History
Originally the CIRS project had baselined the use of one of two existing spare telescopes from the
Voyageur Mid-Infrared Interferometer Spectrometer (MIRIS) as the CIRS flight unit. Extensive testing was
performed by GSFC in conjunction with Applied Optics Associates and ITEK on both the primary mirrors
and the telescopes. However, the MIRIS telescopes' cold performances did not satisfy the CIRS required
cold encircled energy of 80% in a460 micron spot diameter. The warm performance of the best of the two
units was 80% in 880 microns. The spot size grew to 80% in 2200 microns at the operating temperature
of 170 Kelvin _. This was the first cryogenic image quality data obtained for the MIRIS units. Previously,
the telescopes were simply cooled to 135 Kelvin and the absence of change in the warm image spot size
was used as performance acceptance criteria 8.
The degraded telescope performance at cold temperatures is attributed almost entirely to the primary mirror. The
contributing factors include anisotropy of the I70 vacuum hot pressed (VHP) beryllium, bimetallic stresses from
the nickel plating, uneven nickel thickness, and an asymmetric lightweighting design. The anisotropy of the I70
VHP beryllium ranges is approximately 4%. The nickel thickness was estimated (based on a third unit measured
and used on the Mars Observer Laser Altimeter project) to vary from 64 to 97 micorns (0.0025 to 0.0038 inches)
on the front and 122 to 147 microns (0.0048 to 0.0058 inches) on the back. The MIRIS lightweighting pattern is
shown in Figure 3.
HavingdeterminethdattheMIRIStelescopceosuldnotbeusedforflight,anewtelescopweasbaselined.
(OneoftheexistingMIRISunitswasusedfortheCIRSengineerinmgodel.E) mphaswisasplacedon
redesignitnhgeprimarymirror.
Figure 3: MIRIS Primary Mirror Lightweighting Design
3.0 Telescope Design
For the CIRS flight telescope, we retained the same optical prescription as the MIRIS unit but redesigned
the opto-mechanical properties of the telescope. In order to minimize cost, the redesign was performed at
GSFC, the fabrication was contracted out with the final assembly and cryogenic image quality tests done at
GSFC in a test facility built for this eflk)rt. We felt that by eliminating the factors contributing to the
distortion of the primary mirror at 170 Kelvin we could, with low risk, accept delivery of components built
to spec, cold test the assembly in-house, and have a telescope which met our requirements.
The optical prescription was that of a 3meter focal length Cassegrain design and consisted of aparabolic
primary with a vertex focal length of 35.54 cm (14 inches), outside clear aperture of 0.508 meters (20
inches) and an inside clcar apcrturc of 14.22 cm (5.6 inches). Thc sccondary mirror is a convex hyperboloid
having a vertex radius of curvature of I1.5cm (4.528 inches) and an outside clear aperture diameter of 7.96
cm (3.133 inches) and an inside clear aperture of 2.54 cm (1.0 inch). Both mirrors had a 40 _, rms
microroughness requirement. The telescope assembly initially had a back l'_cal length requirement of
11.476 cm +.051cm (4.518 + .020 inches) but this was modified to 11.605 cm - 0.013 + 0.025 cm (4.569
-0.005 + 0.010 inches) based on as-built alignment and test data of the collimating optics 9. The boresight
requiremewnat sthattheblurcirclecentebrelocatewdithina183micron10.007in2ch)radiusofthe
primarymirroropticaalxis.
Thetelesc_pdeesignutilizesthenewe1r70hotisostatipcresseIdHIP)berylliums,ymmetrilcightweighting
designa,ndbareberylliumpolishinagndcoatintgechnologyT.ileHIPberylliumhas0.4%anisotropy
comparewdith4%forthe170VHPberylliumT.helightweightinpgatterndesignebdyBillHaydeannd
PamGeneriocfGFSCisshowninFigure4andshowsthetri-celsl ymmetrwyhichisnotpresenintthe
MIRISunit.Finallyt,hebimetalliecffectswereeliminatebdyutilizingnewcapabilitieinsbareberyllium
polishinagndprotecteedv,aporatgeodldcoating..
_
C_ /47, C1"_'c)
Figure 4: CIRS Primary Mirror Lightweighting Design
The secondary mirror did not undergo redesign but was fabricated out of the new I70H optical grade
beryllium. The mirror was then nickel plated with an overcoat of gold plating.
All the beryllium was purchased from Brush-Wellman and machined by Speedring. The primary mirror was
ground and polished by OCA Applied Optics and coated by Denton Vacuum. The secondary mirror was
also ground and polished by OCA who also applied the electroless nickel plating. The final gold plating
was done at GSFC.
An integral part of the telescope assembly used for both the engineering and flight unit telescopes is the
transition tube which provides a stress free interface between the beryllium primary mirror and the
aluminum optical bench.
Figure 5: CIRS Aluminum to Beryllium Transition Tube
4.0 Warm Assembly and Alignment
The telescope assembly was received as individual components: the primary mirror (PML secondary mirror
(SM), the secondary mirror support tube (SMST), secondary mirr_r shims, and transition tube. The parts
wereassembleadli,gneda,ndtesteadtroomtemperatutrhee,nintegrateindtothecoldchambeforrtestingat
170K.
Upondeliveryandassembolyfthetelescopceomponenwtse,firstmeasurethdebackfocapl ositioonfthe
systemA. setofsecondamryirrorshimswerelocatebdetweetnhesecondamryirrorsupportutbeandthe
secondamryirrormountT. heseberylliumshimswerelappetdoadjustthebackfocallengthofthe
telescope.
Thetransitiotnubep,rimarymirrora, ndtheSMSTwerefirstboltedtogethevriaeightboltsaroundthe
centeorftheprimarymirror.Aberylliumcenterinsgleevwehichrestedinthetransitiotnubewasusedto
centetrheSMSTontothetransitiotnubeT. hisassembwlyasmounteodntoaoneinchthickaluminum
platepolishefdlatonbothsidewsiththesurfacepsarallelT.heassembwlyasthenmounteodnagimbal
mounitnfrontofa24inchcollimatehdelium-neobneam.
WenowalignedtheprimarymirroropticaalxistothecollimatebdeamasshowninFigure6.Todothis,
weplacedatheodolitbeehindtheprimarymirrorlookingthrougihtscentrahloleatthecollimatebdeam.
A
crosshair reticle was then installed into the primary mirror's central hole and the theodolite was translated
and tilted until we could scan through its focus to view the crosshairs and the image formed by the primary
mirror on the same axis. The theodolite was then focused at infinity to view the axis of the collimated
beam. The primary mirror assembly was now tilted to place its optical axis coincident with the collimated
beam. When this occurred, the focus of the theodolite could be adjusted and the image, the crosshairs and
the collimated beam would all lie along the same direction. To verify that the image and the crosshairs
defined the PM optical axis, a fiber light was placed at the on-axis radius of curvature such that the light
was returned back on itself. This was observed to be on the same axis as that defined by the center of the
PM central hole and the PM image of the collimated beam.
Figure 6: Alignment of PM optical axis _ a-_ (._(J
As an alignment aid, we then characterized the pointing direction of the aluminum interface plate with
respect to the PM optical axis. This gave the offset of the transition tube interface to the optical axis of the
PM which was 15 arcseconds in azimuth (horizontal) and 8 arc seconds in elevation (vertical). We could
then easily refer to the flange pointing direction to align the PM to the collimated beam without going
through the cumbersome procedure described in the preceding paragraph.
The secondary mirror and shims were then installed and the back focal position was measured. The SM
decenter position was first adjusted to minimize the image size and off axis misalignment as viewed through
a microscope behind the telescope. The microscope was mounted on a three-axis translator and a reticle was
mounted to the transition tube as shown in Figure 7. Two different reticles were used. The first reticle was
used to measure decenter. This reticle was on aglass plate precisely centered with respect to the central hole
of the primary mirror. We first focused the microscope on the center of the reticle and its horizontal and
vertical micrometer readings were recorded. We then translated the microscope to the center of the image.
The difference between the two sets of micrometer readings was the boresight error. The glass reticle was
then replaced with a thin transparency whose focus position with respect to the back surface of the primary
mirror had been well characterized. Again, we first focused the microscope on the transparency and
translated back to focus on the image. These two focus positions gave us the back focal distance (measured
from the back flange of primary mirror). The focus error was divided by the longitudinal magnification of
73.44 to determine the amount of material to remove from the secondary mirror shims.
Figure 7: Measurement of back focal position
The SM shims were gradually lapped (removing only 2.5 to 5 microns each time) to slowly bring the
image to within the envelope for the back focal distance. When we achieved a back focal length of I1.615
cm (4.573 inches), we measured the boresight error and image size, also using the microscope. The 100%
image size was approximately 430 microns in diameter and was measured by translating the microscope
from one edge of the image to the other.
The boresight error measured was 48 arcseconds inelevation (0.69 mm vertical decenter down) and 12
arcseconds in azimuth (0.18 mm horizontal decenter to the right). The vertical boresight error value
exceeded the boresight requirement that the image fall within 0.183 mm of the optical axis. To
accommodate this boresight error we first verified that tilting the telescope by the above amounts placed the
image on axis. We then verified that no noticeable change in image resulted. With this information we
were able to overcome the boresight error. We did this by tilting the CIRS instrument, after the telescope
was integrated, to center the telescope image on the MIR field and shimming the focal planes to eliminate
any vignetting.
To assist our alignment of the telescope to the collimated beam once the telescope had been secured in the
cold chamber, we characterized the pointing direction of the back of the SM to the telescope optical axis.
This offset value was 15arcseconds in azimuth and 4 arcseconds in elevation. The assembled telescope is
shown in Figure 8without harnessing and blankets.
Figure 8: Side view of the assembled telescope
5.0 Cryogenic Test Chamber
To additionally reduce the cost of the flight telescope, the cryogenic optical performance verification testing
was brought in-house rather than contracted out. The facility, developed specifically for this test, utilized a
chamber already on-hand from the Cosmic Background Explorer project. The chamber dimensions were
modified to accommodate the telescope interface. We then integrated thermal instrumentation, shrouds and
blanketing, CCD camera, and translation stages.
The chamber was originally designed for vertical use which would not suit our purposes. Therefore, we
first built a dolly to mount it horizontally. The chamber's inside diameter of the nitrogen shroud was 0.86
meters (34 inche)s which easily accommodated the half-meter diameter telescope. However, an extension to
the 1meter inside length of the dewar was required. This extension was also 1meter in length and was
mounted on a separate dolly. A 0.61 meter (24 inch) diameter window was mounted in the extension as
were two feedthroughs for the tip and tilt alignment rods. Figure 9 shows the outside configuration of the
modified chamber.
Figure 9: Cryogenic Chamber
The telescope interface was via a second polished aluminum plate which also bolted onto the transition
tube. This plate contained copper tubing through which liquid nitrogen flowed and cooled the telescope
assembly. The telescope is then bolted via the cold plate onto a gimbal mount which connected to the tip
and tilt rods. The gimbal mount is secured to a large platform on rails and which can roll out of the
chamber onto a mating cart allowing easy installation and handling.
A 24 inch diameter, 22 inch long shroud painted thermally black was hard mounted to the large platform.
The gimbal mount is positioned behind the shroud such that thc SMST fits inside and the PM edge is flush
with the end of the shroud as shown in Figure 10. The shroud provides radiative cooling of the SM and
reduces thermal gradients. Multi-layer thermal blanketing material covers the back of the primary mirror and
attaches to the shroud. This is required because the camera and translation stages are kept at ambient
temperature. Platimun resistance thermometers are located on the back of the SM, along the length of the
SMST and at multiple positions on the back of the primary mirror. 50 ohm heaters were used in
conjunctiownithaLakeshoCreryotronic3s30Controlletromaintaitnhe170KelvintesttemperaturAe.
frontviewofthetelescopinestalleidnthechambeisrshownin
Figure I I.
Figure 10: The back of the primary mirror mounted onto the cold plate and __
mount inside the shroud prior to attaching the thermal blanketing. _t
Figure 11: Front view of the telescope mounted in the chamber.
A CCD camera was mounted on a three axis translation stage assembly which bolted to the large platform
behind the gimbal mount. DynaOptic Motion motors were used to drive each translation stage. The motors
were characterized to have less than 9counts of backlash. Each motor was calibrated prior to the test. The
focus motor has 1369 counts/mm and the vertical and horizontal motors have 9690 counts/mm. The
detector used for this test isa 1024 x 1024 array of 18 micron square pixels manufactured by Texas
Instruments of Japan. It is a virtual phase charge coupled device running at 250k pixels per second with the
drive electronics and software provided by the GSFC's Detector Systems Branch Code 553. Integration
time is adjustable from 0.5 seconds to 1000 seconds and was limited by a large shutter used block the 4
inch collimated Zygo beam prior to the 24 inch beam expander. The camera, stages and motors are
thermally isolated from the chamber and operate at ambient temperature and are shown in Figure 12.
L
Figure 12: Side view of the CCD camera mounted on translation stages. _._6_#-
6.0 Cryogenic Test
After placing the back focal point at its correct position and characterizing the pointing direction of the
telescope boresight with respect to the back flange of the secondary mirror, we integrated the telescope into
the cold chamber.
Once assembled into the chamber and the chamber extension secured, the tip and tilt rods were used to align
the telescope optical axis to the collimated beam. A theodolite looking at the return from the back of the
SM was used to align the boresight of the telescope to the collimated beam to within 5 arcseconds. The test
configuration is shown in Figure 13.
/ f D
Figure 13 -# Alignment of the telescope to the collimated beam. _,_¢ _ _d "
We were now ready to acquire ambient images. The first step in the image acquisition procedure was to
darken the room, close the shutter, and record background image files. Each image saved is actually the
average of four acquired frames. For each through focus series, a background image file was created and
saved. This background file was specified as the reference frame which was automatically subtracted from
each telescope image file. After obtaining the background image files, the shutter was activated and the
camera was translated to the telescope focus and test images were acquired to check for saturation. The beam
attenuation and shutter speed were adjusted accordingly. Several telescope images were acquired at motor
increments of 600 counts (438 microns) over +2.36 mm range about focus.
A turbo pump was used to evacuate the chamber and the telescope was cooled to 170 Kelvin at a rate of 20
degrees per hour as measured by the sensor located on the beryllium side of the transition tube. At this
location, the temperature was consistently mid-way between that of the PM (168K) and that of the SM
(172K). After the telescope has soaked for one hour, background and telescope image files are again
acquired as described above.
In addition to verifying the cold encircled energy, the stability of the telescope back focal point was verified.
First, the camera and telescope were removed and the chamber was instrumented with dial indicators on the
translation stages with their points touching the telescope aluminum mounting plate. The readouts of the
stages were recorded for specified indicator readings. Vacuum was pulled in the chamber and the mounting
plate was cooled to 170 Kelvin. The stages were then moved out and back to repeat the indicator readings
and the stage readouts recorded. The difference in stage counts provided us with the motion of the camera
withrespectotthetelescopmeountinpglatebetweewnarmandcold.Secondlwy,erecordethdestagelocus
positionreadouftosrbesftocuswarmandcoldThesevaluewserecombinetodgiveanetimagemotion
withrespectotthetelescopineterface.
TheboresighsttabilitywasalsoverifiedT. hiswasdonedetermininthgeccntroidlocationosfthewarmand
coldimagesT.hecentroipdositioncsouldbedirectlycomparebdecausper,iortoeachsetofthrougfhocus
imagesth,etelescopweasalignetdothecollimatebdeam.
6.1 Results/Summary
The encircled energy calculations were done after afull cycle of data was acquired. The encircled energy was
calculated using aMatLab program. A threshold value was set below which the pixel would not be used in
the sum. This was required because the camera exhibited astreaking both warm and cold which could not
be contributed to saturation. The total energy was then summed, the centroid found and the 80% ensquared
energy width about that centroid was calculated. (The 80% cold ensquared energy requirement was 500
microns.) These 80% values were then plotted as a function of focus position and the best focus position
was determined. A near-best focus image for warm and cold are shown in Figure 14.
Figure 14: Near focus telescope images. Left is warm. Right is cold
The warm image size was 288 microns versus a cold image size of 207 microns. These measurements
repeated to within + 1pixel or + 18microns.
The image movement in focus was determined to be 25 microns towards the telescope. This was calculated
after measuring the camera to telescope separation increased by 127 microns between warm and cold. The
camera had to move 152 microns closer to the telescope to be at the best focus at cold temperatures versus
at warm. The cold image thus had a net motion of 25 microns closer to the telescope giving a final back
focal length of 11.613 inches (nominal = 11.605 - .013 + .025 cm).
As seen in the images in Figure 14, some boresight motion occurred. The image moved "up" 9 arcseconds
(7 pixels) giving a final error of 39 arcseconds vertically (elevation). The image moved to the "left" on the
CCD by 23 arcseconds (19 pixels) giving a final error of 35 arcseconds horizontally.
Both the image size and back local position easily meet the required values at cold temperatures. As
mentioned in section 4.0 the boresight error was removed by tilting the instrument to center the telescope
image on the MIR field of view and then shimming the MIR focal plane assembly as needed.
7.0 Acknowledgments
The authors wish to thank Tom French, Lou Worell of NSl/ManTech, John Zaniewski, Dan McHugh, Armando
Morell, Bill Hayden, Marilyn Kraut, Laddawan Ruamsuwan, Pare Generic, Ravi Kaipa of Hughes STX, Dan
Musinski (formerly of NSI, now at Zygo Corp), and Swales Aerospace.
8.0 References